Close

State of the project, March 2022

A project log for PipeCam: Low-Cost Autonomous Underwater Camera

Low cost autonomous underwater camera for long term deployments and exploration

fred-fourieFred Fourie 03/21/2022 at 19:010 Comments

State of the project:

  1. PCB version is currently at V06.04. Code named DORY
  2. V06.04 is a single PCB with only two external peripherals: a power sensor chip and a RTC
  3. The new design ditched the external controller altogether
  4. There is no longer an external SD Card slot
  5. DORY now has two extra switches for toggling devices.
  6. DORY features wifi data download, to make testing easier
  7. A new housing with an external on/off switch was tested and flooded. (more later)
  8. Camera test have revealed that the benefits of the HQ cameras are not THAT great and that the V2 raspberry pi camera still performs really well
  9. DORY has not been in the water yet and is currently undergoing low power testing. (Hence no cool pictures as of yet)

Ditching of the micro-controller:

This is by far the biggest change to the project. The old versions of all relied on an external micro-controller (the Arduino ProMini clone) reading an SD Card to determine the sleep intervals.

This meant that, once deployed, there was no way to change this apart from opening up the housing and getting to the SD Card. This was proven a problem in when working in the field. 

Further more, a way to use the RTC alone on to do the power switching of the pi was devised. Removing the micro-controller would remove the need for an external SD Card (discussed next), reduce the power and reduce the complexity of the project. Ironically, shortly after this design choice was made, I found that the Robodyn Promini boards which I had used for my previous boards has been formally discontinued. 

Removal of the SD card:

Not all SD Cards are created equal. Between brands and individual units, SD Cards differ. This is became very apparent when I started measuring the low-power sleep modes in the previous iterations. Different SD Cards would push the power consumption from a couple of uA to as much as 1mA in some bad cases. While troubleshooting this, I came across this from another inspirational hackaday project:

https://thecavepearlproject.org/2014/09/22/high-sleep-current-problem-solved/

With all of these troubles in mind, I decided to ditch the SD Card.

The Great Silicon Shortage:

It has been incredible to see that usually, easy-to-find components are just globally out of stock. This has pushed the need to redesign the board with less component and towards designing for interchangeable footprints.  

DON'T RUSH IT

Over the last hear I had the opportunity to deploy at sea. I quickly made a new housing and introduced a number of "new" components to the housing. This catastrophically failed within minutes of deployment. Mechanical stuff is not my strong point. So next time I should take my time, test the housing.

Alternatives:

If you have landed on this because you are itching to build your own project, similar to this. I have some good news for you. There are off-the-shelf components that will help you get rolling MUCH faster than waiting for this project. The best example I have found so far is the Witty Pi:

https://www.uugear.com/product/witty-pi-3-realtime-clock-and-power-management-for-raspberry-pi/

The near future:

The new DORY PCB is currently undergoing testing, so far all the peripherals are performing as expected and the PCB design seems sound. 

The next step will be a redesign of the internals: a battery pack and a set of mountings for the PCB and camera.

And the hopefully! Back into the water she goes!

Discussions